IBHM: index-based data structures for 2D and 3D hybrid meshes

Abstract

We propose new topological data structures for the representation of 2D and 3D hybrid meshes, i.e., meshes composed of elements of different types. Hybrid meshes are playing an increasingly important role in all fields of modeling, because elements of different types are frequently considered either because such meshes are easier to construct or because they produce better numerical results. The proposed data structures are designed to achieve a balance between their memory requirements and the time complexity necessary to answer topological queries while accepting cells (elements) of different types. Additionally these data structures are easy to implement and to operate, because they are based on integer arrays and on basic arithmetic rules. A comparison with other existing data structures regarding their memory requirements and of the time complexities for the algorithms to answer general topological queries is also presented. The comparison shows that the overhead required to accept arbitrary cell types is small.